Limit of Detection Calculator

A detection limit answers a statistical question, not a chemical one: at what level does a signal become distinguishable from what a blank would produce by chance. The three routes offered here approach it differently. The blank method adds k standard deviations of repeated blank measurements to the blank mean, giving a threshold in signal units. The calibration method divides k times the residual standard deviation of the regression by the slope, giving a threshold directly in concentration units, which is the form specified in ICH guideline Q2 for pharmaceutical method validation. The signal-to-noise approach compares analyte response against peak-to-peak baseline noise, conventionally 3 to 1 for detection and 10 to 1 for quantitation, and is standard in chromatography where noise is visible on the trace. The factor k is where most confusion arises. Using 3 sigma controls only false positives, at roughly 7 percent one-sided for a normal distribution when the blank standard deviation is well estimated. Controlling false negatives as well requires about 3.3 sigma, which comes from summing two one-sided 95 percent normal quantiles of 1.645, and this is why 3.3 rather than 3 appears in ICH-based work. All three methods share a critical dependency: the standard deviation must be estimated from enough replicates, since with only 5 or 7 blanks the estimate itself is uncertain by tens of percent. Regulatory procedures address this directly, with the United States EPA method detection limit protocol using a Student t multiplier at 99 percent confidence applied to at least seven spiked replicates rather than a fixed factor.

Formal detection limits emerged from spectrochemical analysis in the mid-twentieth century. Heinrich Kaiser, working in Germany on emission spectrography, argued that the smallest detectable amount had to be defined relative to the fluctuation of the blank rather than by an analyst's judgement of whether a line was visible, and proposed the three standard deviation criterion that remains the intuitive core of the concept. The reasoning was formalised and corrected by Lloyd Currie of the United States National Bureau of Standards in a 1968 paper in Analytical Chemistry on limits for qualitative detection and quantitative determination, which distinguished three separate quantities: a critical level for deciding whether a signal is present, a detection limit describing what the method can reliably find, and a determination limit for acceptable quantitative precision. Currie's framework showed that a single three sigma threshold conflates false positive and false negative risk, and that a symmetric treatment of both errors leads to approximately 3.29 sigma, the origin of the 3.3 factor now in routine use. IUPAC issued recommendations on detection and quantification capabilities in 1995 based largely on that treatment. Parallel regulatory codification followed different paths: the United States EPA adopted a method detection limit procedure derived from work published by Glaser and colleagues in 1981, written into 40 CFR Part 136, while the International Council for Harmonisation guideline Q2 fixed the slope-and-residual-deviation formulation for pharmaceutical analysis.